CycleGAN训练教程

在输入图像和输出图像之间，使用对齐图像对的训练集。但是，对于许多任务，配对训练数据将不可用。我们提出一个学习从源翻译图像的方法域 X 到目标域 Y，在没有配对的情况下例子。我们的目标是学习映射 G ： X → Y使得来自 G（X）的图像分布与使用对抗性损失的分布 Y 无法区分。由于此映射的约束非常不足，因此我们将其与逆映射 F ： Y → X ，并引入强制执行 F（G（X）） ≈ X 的循环一致性损失（反之亦然）。在几个任务上呈现了定性结果不存在配对训练数据的地方，包括集合风格转移、对象变形、季节转移、照片增强等定量比较。

CycleGAN原理

（a）我们的模型包含两个映射函数 G ： X → Y 和 F ： Y → X，以及相关的对抗函数
鉴别器 DY 和 DX。DY 鼓励 G 将 X 转换为与域 Y 无法区分的输出，反之亦然
用于 DX 和 F。为了进一步规范映射，我们引入了两个循环一致性损失函数，即如果
我们从一个领域转换到另一个领域，然后再转换回来：（b）前向损失函数：x → G（x） → F（G（x）） ≈ x，以及（c）后向损失函数：y → F（y） → G（F（y）） ≈ y

CycleGAN的应用场景

训练过程

（以马<——>斑马的训练过程为例）

代码内容

环境配置

#操作命令：
pip install -r requirements.txt

requirements.txt内容如下：

torch>=1.4.0
torchvision>=0.5.0
dominate>=2.4.0
visdom>=0.1.8.8
wandb

torch和torchvision安装建议直接官网配置：PyTorch

预训练权重下载

进入到根目录pytorch-CycleGAN-and-pix2pix-master/下面：

运行命令：

bash ./scripts/download_cyclegan_model.sh horse2zebra

下载到.//checkpoints/horse2zebra_pretrained目录下面

也可以网页直接下载:Index of /cyclegan/pretrained_models (berkeley.edu)

下载训练数据

运行命令：

bash ./datasets/download_cyclegan_dataset.sh maps

下载到：./datasets/horse2zebra目录下面

也可以直接网页下载：打开./datasets路径下的download_cyclegan_dataset文件查看到数据下载路径为：Index of /cyclegan/datasets

参数设置

./options目录下

基础参数配置：base_options.py文件

    def initialize(self, parser):
        """Define the common options that are used in both training and test."""
        # basic parameters
        parser.add_argument('--dataroot', required=True, help='path to images (should have subfolders trainA, trainB, valA, valB, etc)')
        parser.add_argument('--name', type=str, default='experiment_name', help='name of the experiment. It decides where to store samples and models')
        parser.add_argument('--gpu_ids', type=str, default='0,5,6,7', help='gpu ids: e.g. 0  0,1,2, 0,2. use -1 for CPU')
        parser.add_argument('--checkpoints_dir', type=str, default='./checkpoints', help='models are saved here')
        # model parameters
        parser.add_argument('--model', type=str, default='cycle_gan', help='chooses which model to use. [cycle_gan | pix2pix | test | colorization]')
        parser.add_argument('--input_nc', type=int, default=3, help='# of input image channels: 3 for RGB and 1 for grayscale')
        parser.add_argument('--output_nc', type=int, default=3, help='# of output image channels: 3 for RGB and 1 for grayscale')
        parser.add_argument('--ngf', type=int, default=64, help='# of gen filters in the last conv layer')
        parser.add_argument('--ndf', type=int, default=64, help='# of discrim filters in the first conv layer')
        parser.add_argument('--netD', type=str, default='basic', help='specify discriminator architecture [basic | n_layers | pixel]. The basic model is a 70x70 PatchGAN. n_layers allows you to specify the layers in the discriminator')
        parser.add_argument('--netG', type=str, default='resnet_9blocks', help='specify generator architecture [resnet_9blocks | resnet_6blocks | unet_256 | unet_128]')
        parser.add_argument('--n_layers_D', type=int, default=3, help='only used if netD==n_layers')
        parser.add_argument('--norm', type=str, default='instance', help='instance normalization or batch normalization [instance | batch | none]')
        parser.add_argument('--init_type', type=str, default='normal', help='network initialization [normal | xavier | kaiming | orthogonal]')
        parser.add_argument('--init_gain', type=float, default=0.02, help='scaling factor for normal, xavier and orthogonal.')
        parser.add_argument('--no_dropout', action='store_true', help='no dropout for the generator')
        # dataset parameters
        parser.add_argument('--dataset_mode', type=str, default='unaligned', help='chooses how datasets are loaded. [unaligned | aligned | single | colorization]')
        parser.add_argument('--direction', type=str, default='AtoB', help='AtoB or BtoA')
        parser.add_argument('--serial_batches', action='store_true', help='if true, takes images in order to make batches, otherwise takes them randomly')
        parser.add_argument('--num_threads', default=4, type=int, help='# threads for loading data')
        parser.add_argument('--batch_size', type=int, default=32, help='input batch size')
        parser.add_argument('--load_size', type=int, default=286, help='scale images to this size')
        parser.add_argument('--crop_size', type=int, default=256, help='then crop to this size')
        parser.add_argument('--max_dataset_size', type=int, default=float("inf"), help='Maximum number of samples allowed per dataset. If the dataset directory contains more than max_dataset_size, only a subset is loaded.')
        parser.add_argument('--preprocess', type=str, default='resize_and_crop', help='scaling and cropping of images at load time [resize_and_crop | crop | scale_width | scale_width_and_crop | none]')
        parser.add_argument('--no_flip', action='store_true', help='if specified, do not flip the images for data augmentation')
        parser.add_argument('--display_winsize', type=int, default=256, help='display window size for both visdom and HTML')
        # additional parameters
        parser.add_argument('--epoch', type=str, default='latest', help='which epoch to load? set to latest to use latest cached model')
        parser.add_argument('--load_iter', type=int, default='0', help='which iteration to load? if load_iter > 0, the code will load models by iter_[load_iter]; otherwise, the code will load models by [epoch]')
        parser.add_argument('--verbose', action='store_true', help='if specified, print more debugging information')
        parser.add_argument('--suffix', default='', type=str, help='customized suffix: opt.name = opt.name + suffix: e.g., {model}_{netG}_size{load_size}')
        self.initialized = True
        return parser

训练参数设置：train_options.py


    def initialize(self, parser):
        parser = BaseOptions.initialize(self, parser)
        # visdom and HTML visualization parameters
        parser.add_argument('--display_freq', type=int, default=400, help='frequency of showing training results on screen')
        parser.add_argument('--display_ncols', type=int, default=4, help='if positive, display all images in a single visdom web panel with certain number of images per row.')
        parser.add_argument('--display_id', type=int, default=1, help='window id of the web display')
        parser.add_argument('--display_server', type=str, default="http://localhost", help='visdom server of the web display')
        parser.add_argument('--display_env', type=str, default='main', help='visdom display environment name (default is "main")')
        parser.add_argument('--display_port', type=int, default=8097, help='visdom port of the web display')
        parser.add_argument('--update_html_freq', type=int, default=1000, help='frequency of saving training results to html')
        parser.add_argument('--print_freq', type=int, default=100, help='frequency of showing training results on console')
        parser.add_argument('--no_html', action='store_true', help='do not save intermediate training results to [opt.checkpoints_dir]/[opt.name]/web/')
        # network saving and loading parameters
        parser.add_argument('--save_latest_freq', type=int, default=5000, help='frequency of saving the latest results')
        parser.add_argument('--save_epoch_freq', type=int, default=5, help='frequency of saving checkpoints at the end of epochs')
        parser.add_argument('--save_by_iter', action='store_true', help='whether saves model by iteration')
        parser.add_argument('--continue_train', action='store_true', help='continue training: load the latest model')
        parser.add_argument('--epoch_count', type=int, default=1, help='the starting epoch count, we save the model by <epoch_count>, <epoch_count>+<save_latest_freq>, ...')
        parser.add_argument('--phase', type=str, default='train', help='train, val, test, etc')
        # training parameters                     ##总共epoch niter+niter_decay=200轮
        parser.add_argument('--niter', type=int, default=100, help='# of iter at starting learning rate')
        parser.add_argument('--niter_decay', type=int, default=100, help='# of iter to linearly decay learning rate to zero')
        parser.add_argument('--beta1', type=float, default=0.5, help='momentum term of adam')
        parser.add_argument('--lr', type=float, default=0.0002, help='initial learning rate for adam')
        parser.add_argument('--gan_mode', type=str, default='lsgan', help='the type of GAN objective. [vanilla| lsgan | wgangp]. vanilla GAN loss is the cross-entropy objective used in the original GAN paper.')
        parser.add_argument('--pool_size', type=int, default=50, help='the size of image buffer that stores previously generated images')
        parser.add_argument('--lr_policy', type=str, default='linear', help='learning rate policy. [linear | step | plateau | cosine]')
        parser.add_argument('--lr_decay_iters', type=int, default=50, help='multiply by a gamma every lr_decay_iters iterations')

测试参数配置：test_options.py

    def initialize(self, parser):
        parser = BaseOptions.initialize(self, parser)  # define shared options
        parser.add_argument('--ntest', type=int, default=float("inf"), help='# of test examples.')
        parser.add_argument('--results_dir', type=str, default='./results/', help='saves results here.')
        parser.add_argument('--aspect_ratio', type=float, default=1.0, help='aspect ratio of result images')
        parser.add_argument('--phase', type=str, default='test', help='train, val, test, etc')
        # Dropout and Batchnorm has different behavioir during training and test.
        parser.add_argument('--eval', action='store_true', help='use eval mode during test time.')
        parser.add_argument('--num_test', type=int, default=50, help='how many test images to run')
        # rewrite devalue values
        parser.set_defaults(model='test')
        # To avoid cropping, the load_size should be the same as crop_size
        parser.set_defaults(load_size=parser.get_default('crop_size'))

训练操作

命令:

python train.py --dataroot ./datasets/horse2zebra --name maps_cyclegan --model cycle_gan

训练界面

训练结果

保存在：./checkpoints/maps_cyclegan文件下面

测试操作

测试命令

python test.py --dataroot ./datasets/maps --name maps_cyclegan --model cycle_gan

测试结果

在路径.results\maps_cyclegan\test_latest_maps下可以看到测试结果

real为原图，fake为具有原图风格的假图，rec为根据假图复原原图的图。

项目下载链接

CycleGAN概述